Application of bioinformatics algorithms to define the most important protein features contribute to GABA receptors diversity

Abstract

The importance of GABA receptors in the modulation of transmitter release and the late inhibitory postsynaptic potential and their ubiquitous distribution within the CNS promise a good deal as targets for many medical and pharmacological interventions. GABA interacts with various receptors types as A, B, C and various subtypes receptors. Various methods have been employed to distinguish between different GABA receptors' identity and here we have used various bioinformatics algorithms such as Rule Induction, Tree Induction, Attribute Weighting and clustering models to find out the most important protein features in each GABA receptors. More than 900 protein features for sequences all known GABA receptors were compared and the results showed in majority of models employed in this study, the frequencies and the counts of dipeptides play major roles in forming various types of GABA receptors. It has also shown the performance and the accuracies of the models employed here were generally high enough (except for one) confirming this approach can be used to study the structural difference between types and subtypes of GABA receptors. The new finding will be carefully analyzed and will be enclosed in camera ready paper.